Interactive Transcript
0:02
Hello, and welcome to Noon Conference hosted by Modality.
0:05
Noon Conference connects the global radiology community through free, live
0:09
educational webinars that are accessible for all and is an opportunity to learn
0:12
alongside top radiologists from around the world.
0:16
Today, we're honored to welcome Dr.
0:17
Prashant Nagpal for a lecture entitled "Mastering Cardiac CT: Essential
0:22
Fundamentals and Emerging Clinical Frontiers." Dr.
0:25
Nagpal completed his radiology residency at the University of Iowa and his
0:30
cardiovascular imaging fellowship at Brigham and Women's Hospital, Harvard Medical
0:33
School. He currently leads the section of cardiovascular imaging at the University
0:37
of Wisconsin-Madison, specializing in advanced cardiac and vascular imaging.
0:43
At the end of the lecture, please join him in a Q&A session where he will address
0:46
questions you may have on today's topic.
0:49
Please remember to use that Q&A feature to submit your questions so we can get to
0:52
as many as we can before our time is up.
0:55
With that, we are ready to begin today's lecture. Dr.
0:57
Nagpal, please take it from here.
1:00
Thank you. Thank you so much. First of all, welcome, everybody.
1:03
Thank you for attending this noon lecture about cardiac CT. Okay.
1:08
My goal is to touch upon two spectrums of cardiac CTs, knowing that the learners
1:15
are from different backgrounds, they all have different areas of interest.
1:19
So I'm going to teach a little bit about the basics of cardiac CTs and talk about
1:24
how we acquire and what are the basics that we must keep in mind, and also touch
1:28
upon where the field is going. Obviously, it's growing so vast and in so many
1:34
areas, I cannot do that all in one lecture, but I'm just going to give a brief
1:39
trailer of what's coming. With that, let's dive in.
1:45
First of all, here are my disclosures.
1:48
And along with that, I certainly also
1:54
work with Modality for these educational initiatives.
2:00
So the objectives of this lecture today is, one, to understand how cardiac CT
2:06
is acquired and understand basic protocol and acquisition techniques.
2:12
This is one of the basics, especially for early learners, those who are just
2:17
starting to learn how cardiac CT is done.
2:22
We'll briefly talk about some emerging role of advanced applications for coronary
2:27
CT, as well as discuss a little bit of data on why coronary CT is such an important
2:33
test. Beyond that, we will discuss some case examples to show the role of cardiac
2:39
CT for coronary evaluation, as well as some non-coronary cardiac evaluation.
2:45
With that, first, let's dive in in the basics and imaging principles.
2:51
So the way I think of cardiac CT is the journey of a cardiac CT, and
2:57
again, I call it a journey because there are multiple steps involved.
3:01
So journey obviously initially starts from an order.
3:06
I'm not going to talk about the order part, but as an imager, it
3:11
starts at the time of patient preparation and then involves image acquisition.
3:17
Then once the images are acquired, it goes on to image display, understanding
3:23
artifacts, as well as understanding common and challenging scenario that we
3:28
encounter in our day-to-day practice.
3:31
The common ones are high heart rate, arrhythmias, obesity, stents, and other
3:37
metal things in the heart, like pacemaker and ICDs.
3:42
I'll try to touch upon most of these.
3:45
Obviously, this is a one-hour lecture, I cannot do that all in this one hour, but
3:51
I'll try my best to touch upon things.
3:55
So, as I said, our journey starts when the patient arrives.
3:59
So in terms of how do we take care of the patient, the basic instructions that we
4:04
provide to patients are they're supposed to hydrate normally.
4:08
We ask them to have water up to the time of scan, and obviously encourage water
4:13
after. After this,
4:16
we also encourage normal diet. We don't ask people to have overnight fast or
4:21
anything for scan itself. One preference is to have no food for three to four
4:26
hours. But more important than just food, we really focus on things that can
4:32
affect patient preparation in terms of no caffeine for 20 to 24 hours,
4:38
including decaffeinated drinks.
4:40
This is because we want to have slow, steady heart rates.
4:43
Similar logic applies to exercising.
4:48
We basically are aiming for a slow, steady heart rate.
4:51
We encourage people and ask people to continue their usual cardiac medications like
4:56
beta blockers, but certain medications like phosphodiesterase inhibitors, like
5:01
sildenafil or tadalafil, should be discontinued 48 hours prior.
5:08
Obviously, there are other checks that happen before even the patient comes.
5:13
We make sure we look for prior contrast reactions, the type and severity.
5:19
We ensure patient's renal function is okay, and obviously look for pregnancy status
5:24
as well as any metal in or near the heart which can affect the
5:30
reconstruction protocol. Then beyond that,
5:34
this is again, a local protocol. We use a single source scanner.
5:39
So single source, and I'll come to that why this difference is important.
5:43
Single source means that there is one X-ray source.
5:46
And for those scanners, heart rate control is really critical.
5:51
So since we are a single source site, we aim for less than 60 beats per minute
5:56
heart rate, and we primarily rely on metoprolol for these.
6:01
So we use, depending on the heart rate, metoprolol oral one hour before the scan.
6:07
We sometimes use IV top-ups as needed.
6:10
We also use ivabradine or diltiazem as indicated.
6:16
And of course, knowing these contraindications for these medications, like
6:22
beta blockers with high rate AV block, severe reactive airway, which is probably a
6:27
relative contraindication and
6:30
low blood pressure. And same, we always use sublingual
6:35
nitroglycerin when we can. When I say always, it is obviously making sure that
6:41
patient does not have a contraindication to use of
6:45
these medications which are listed here.
6:48
Our goal is to use at least two tablets, that is 0.4
6:54
milligram, two tablets, that is 0.8, but obviously we give it only when it is
6:59
stable, especially when the systolic blood pressure is more than 110.
7:04
At less than 100 we hold, and between 100 and 110 we give 0.4 milligram.
7:09
Important thing about nitroglycerin that we give sublingual, you do not scan
7:15
right away. You have to wait three to five minutes.
7:18
So what we do is that we basically have the patient on the table, give
7:22
nitroglycerin, and then start planning the protocol and calcium score and then
7:27
contrast administration. That gives us three to five minutes timeframe without
7:31
doing it outside the CT entry. Now, obviously, you have now
7:38
kind of prepared the patient for the scan. You have explained the scan.
7:42
Then comes the part of you are going to inject contrast.
7:48
Now, when we inject contrast, understanding the injection
7:53
protocols is important.
7:56
Injection protocols are either monophasic, in which as you see, all we do is that
8:02
we give contrast. No saline after that. We just inject contrast and we scan.
8:07
This is, I would say, the least commonly or almost non-existing
8:13
contrast protocol. I do not think we use it in any patient.
8:18
The common protocols that we use, biphasic protocol is actually for coronary
8:23
imaging, the most common protocol that we use.
8:26
We basically give contrast followed by saline, and this is kind of our workhorse
8:33
for imaging. Then, there is a triphasic protocol that we also use
8:39
very commonly in cardiac imaging, not for coronary, but for other cardiac imaging,
8:46
in which we give contrast, and then we give a contrast saline mix followed by pure
8:50
saline. This is basically whenever you want to have right and left heart together.
8:56
So for us, there are certain indications.
8:59
So for example, if- You want to have
9:04
intracardiac thrombus evaluation, in which you want to see thrombi in any side of
9:09
the heart. We also use this extensively in our congenital heart disease
9:15
patients because you want to have both pulmonary artery and aorta
9:21
evaluated along with coronaries in this population.
9:25
There is another alternative to having both right and left heart is, very
9:30
certain injectors are capable of doing multiple small splits.
9:35
It's not very common, and it's only injection specific.
9:39
Interestingly, in our practice, we have that injector and we have very good results
9:44
with that. But for you to think of is, this is how
9:50
I understand that for coronary imaging, you will be using biphasic for most part.
9:56
And if you need both left and right heart, you will be using triphasic protocol.
10:01
I think that should kind of simplify the understanding, and monophasic is rarely
10:07
ever the correct answer. In my experience, we almost never use it.
10:13
Saline flush is very important, and
10:17
it pushes the contrast from the veins into the central circulation, and it also
10:21
clears the contrast out of the SVC and right atrium.
10:25
And for coronary imaging, it is very, very important because otherwise you cannot
10:29
see your right coronary artery very well if you have dense contrast in the right
10:33
atrium.
10:35
And I'll show you an example of that, too.
10:37
So this is kind of how our injectors look for those who do not have much experience
10:43
with coronary imaging yet, or any cardiac imaging yet.
10:47
And basically, there are two heads, one in one we fill saline and in one we fill
10:52
contrast. And the screen
10:56
frequently looks like this. If you see, this is an example, if somebody asks you
11:01
what kind of contrast bolus is this? This is a triphasic contrast bolus.
11:05
You have given first contrast, so green is basically contrast, contrast, then
11:10
contrast saline mix, and then only saline.
11:13
So this is a typical example of a triphasic contrast.
11:17
Now, this image shows how a monophasic contrast looks like.
11:23
There is very dense contrast coming in the SVC right atrium.
11:27
And as you can imagine, if you were doing this for coronary imaging, even though
11:31
the left heart assessment may be okay, the right coronary artery is completely
11:36
obscured by artifacts, which we don't want.
11:40
So for that, this is the kind of image we really like to have for coronary imaging.
11:45
That's why I said,
11:47
before coronary imaging, what we want is a biphasic protocol.
11:50
That means we gave contrast, and that contrast is in the left heart, and then we
11:54
gave saline, which kind of washed everything out of the right heart.
11:58
There is still a little bit left, but again, doesn't matter.
12:02
In this area, even if it is there, it is not dense enough to cause streaks, and it
12:08
gives you a very good imaging for the right heart.
12:12
And this contrast setting is also very good to look for very subtle shunts like
12:17
PFO, in which you can see contrast jet going from left to right, sometimes
12:22
you'll see that. So this is our most preferred and most widely used contrast
12:28
bolus. This is the triphasic contrast bolus.
12:32
As I said, we sometimes use it depending on the indication of the study.
12:36
The most common study in which we use is basically congenital imaging.
12:41
So just to give you a perspective, this is a biphasic contrast.
12:46
This is how the images, when you scroll through, look like.
12:48
Beautiful contrast in the left heart and almost nothing in the right heart because
12:53
we gave saline. And the biphasic doesn't mean you cannot do right heart
12:59
imaging with this. This is also a biphasic contrast, but timed for right heart
13:04
evaluation. So you can use biphasic contrast in either ventricle.
13:10
Just the way it helps in the right heart is it kind of gives
13:16
great timing. And this is our triphasic contrast, in which we gave it and now you
13:21
have enhancement of both sides, so you can evaluate diseases on either side.
13:27
So this is about contrast
13:31
administration. Now, let's say you have decided how you are going to inject.
13:37
How do you scan or how do you figure out when to start
13:42
acquiring images once you have given the contrast?
13:46
So the way we do it is called as a scan delay.
13:49
So
13:50
for those who do imaging of other organ systems, like for example, chest imaging or
13:56
abdominal imaging, you probably know that we have a fixed delay
14:02
time. Like for example, for abdominal pelvis routine exams, we wait 60 seconds.
14:08
Or for chest imaging, we wait 60 seconds and then we acquire.
14:11
In cardiovascular imaging, you cannot do that because you
14:18
want to time your acquisition when there is peak enhancement.
14:22
And peak enhancement depends on patient's physiology.
14:26
So what we do is that we basically monitor where the
14:32
contrast is, and based on that monitoring, we look when it is peaking in the
14:37
ascending aorta. Then there is a slight post-threshold delay because we want
14:44
two to four seconds to allow complete filling of the coronary arteries.
14:47
So what we are doing is we are looking at the aorta, looking at the peak of that
14:52
aorta, aortic contrast, and as soon as there is a peak, we basically give breath
14:58
holding instructions like, "Hold your breath." That takes care of that two to four
15:02
seconds for coronaries to fill in, and that is basically how we do it.
15:08
Even in that, there are two methods.
15:10
One is called as a bolus tracking, in which what we do is that we basically take
15:16
an image, and then on that image, we put a ROI.
15:22
And as soon as it crosses a certain threshold, like in this case, it was 100
15:28
Hounsfield unit in the aorta. And as soon as it crossed 100 Hounsfield units in
15:34
the aorta, the scan got triggered.
15:36
So
15:39
basically, this is how it looks. This is an example of bolus tracking.
15:45
And what we do is that depending on what type of enhancement we want, if you want a
15:49
left heart enhancement, you can trigger off either the descending thoracic aorta or
15:54
the ascending aorta. As soon as the contrast crosses, let's say, 100 Hounsfield
15:59
units, there is a slight post-scan delay to give reading instructions, the scan
16:05
would start.
16:08
There is no particular preference in our program.
16:11
We don't like using ascending aorta, even though it is bigger.
16:15
The reason is sometimes the contrast, when it is coming from the SEC, it can cause
16:21
streaks, and those streaks can sometimes falsely trigger the scan
16:26
early. So we personally prefer doing it off the descending thoracic aorta.
16:31
But I've worked at programs that do
16:34
ascending aortic triggering, and somehow it works.
16:37
And again, there's nothing wrong here if your program does that.
16:42
As long as you are not having too many of these cases that are wrongly triggered,
16:46
then you should look into descending thoracic aorta.
16:50
Then obviously, if you want a right heart only scan, just trigger it off the main
16:55
pulmonary artery. We use it for a lot of our scans in which we are looking
17:01
at tricuspid assessment or tricuspid valve implant planning,
17:06
or what we call STVR planning. So we do use them.
17:11
The second type of triggering is called as test bolus.
17:15
This is not used very common. We use it sometimes in patients who have very
17:20
unusual physiology, like in patients with fontane circulation.
17:24
We want to really time it exactly when we do.
17:28
So in this, what we do is that we basically inject contrast, a small amount of
17:32
contrast, and then we see how contrast travels through the body.
17:38
And once we see how it travels through the body, we create these curves.
17:42
So first contrast comes in the pulmonary artery as we inject in the systemic vein,
17:47
and then it comes in the aorta. So we measure the exact time it takes.
17:52
So in this case, it's around 20 seconds it came in the aorta.
17:58
Then we would add a little bit like, for example, four more seconds or three more
18:02
seconds. They are called flush factor or whatever you can call it, time it takes
18:06
for coronaries to fill in. So, on time bolus you get 20,
18:12
you would basically scan at 22 or 23 seconds, and you will get excellent contrast
18:18
enhancement. Obviously, the problem with this is that to get this time bolus graph,
18:24
you have to inject 10 to 20 ml. And it works very well in
18:29
unusual physiology, but then it costs you extra contrast, extra time, and a
18:35
little dose. Also, at least for us, our techs are very comfortable doing bolus
18:41
tracking. It's more efficient. We do so many cases a day.
18:45
We basically practically almost never do test bolus except when we know that we
18:51
have a fontane patient with extreme collateral load coming in, then that's the only
18:56
time we use it. So, okay, so you have decided what type of contrast you are giving,
19:02
how you are going to trigger, what should be your field of view.
19:06
So for normally all coronary CTs, we only acquire the heart, and in that
19:12
acquisition,
19:15
basically we acquire the scout image or a topogram image, and this goes from
19:21
the level of the carina to the base of the heart.
19:23
That's the most common
19:26
Z-axis coverage that we go with. Now, obviously, not every cardiac CT is a simple
19:32
coronary CT. In patients with bypass grafts, we aim to go just above the lung
19:38
APCs because you want to make sure the entire lima to LAD is included.
19:44
You want to assess subclavian origins.
19:47
You want to assess lima patency from the top.
19:51
So that's why you must go higher, and the landmark that we read in literature is
19:57
either you go higher than the lung APCs or you just go to the midcarinal level.
20:01
It depends on how you want to remember.
20:04
So other indications in which we do broader coverage is patients with congenital
20:08
heart disease. As I said, frequently, the question is, we'll get a cardiac CT, but
20:12
the question is, can we make sure patient does not have coarctation or anything
20:17
else going in aorta? So we want to cover the entire heart.
20:20
Also, it's not
20:23
that uncommon to get cardiac CT plus aortic sizing or any other aortic
20:29
pathology together, and in those cases, we cover the whole chest.
20:35
In terms of how we acquire cardiac CT, so obviously,
20:41
heart is a moving organ,
20:43
and you don't want to acquire through motion.
20:47
So we basically
20:49
sync our acquisition with cardiac cycle.
20:53
And we can sync it through three ways.
20:56
Actually, mostly remember it as two ways.
21:01
One is prospective. That means we tell the scanner, "This is the part of the
21:05
cardiac cycle I want acquired." Retrospective is when you acquire throughout the
21:10
cardiac cycle and then later on bin your images in certain parts of cardiac cycle.
21:15
That's how I would think of. And the thing is that prospective or when you tell the
21:21
scanner, "This is exactly I want," that's low dose, but you don't get too much
21:26
data. It has little flexibility. You obviously cannot get cardiac
21:31
function out of it, whereas retrospective gives you a lot more flexibility.
21:37
So prospective ECG triggering is also called as step and shoot mode.
21:41
In this, if you see what happens is you acquire a first slab, then there's stable
21:46
motion. You acquire second slab, there's stable motion, then you acquire third.
21:50
And if you think of your entire acquisition, depending on scanner that you have,
21:56
each slab gets acquired in a certain time period.
22:00
Now, if you have what is called as a volume scanner, that means the scanner that
22:04
can acquire 16 centimeter in one heartbeat, you do not have this problem of
22:10
multiple slabs. But most scanners that are used, they have these small
22:16
slabs. So it's basically acquired across, in this case, four heartbeats.
22:21
So that means you can have what is called a step artifact.
22:25
If patient either has respiratory motion or irregular heart rate, you can
22:31
have some artifacts that come with it.
22:34
So along with that, so this is basically how we basically acquire
22:39
at 40%, and you start, even though this heart rate is so regular, it
22:45
basically was 85 to 78. It's fairly regular heart rate.
22:50
Even then, I can tell all these slabs that were acquired, and I can tell these were
22:55
each one heartbeat. So that's kind of one disadvantage of
23:01
prospectively triggered scan, but at the same time, they come with a lot of
23:05
advantages of lower radiation dose and actually giving you good diagnostic
23:10
information. So I would say other than valve imaging or like TAVR and others,
23:16
I think prospectively triggered scans have become the standard of care for
23:21
cardiac imaging.
23:24
Retrospective imaging, as you can imagine, we discussed, you are acquiring the data
23:29
the whole time. You basically acquire it whole time.
23:31
This is also called as helical. So you basically acquire, now you can acquire with
23:35
full radiation dose, or you can acquire in the desired portion with higher dose and
23:41
then reduce the dose. So that's how we do it.
23:44
And this is where we are changing the dose.
23:48
So in certain areas, you see excellent quality, and then in certain, the image
23:51
quality is relatively poor. This is called as dose modulation.
23:55
Whereas sometimes we acquire full retrospective.
23:58
This is a patient with TAVR who was having increasing gradients.
24:01
So we acquired like a full cardiac CT, and in this we had
24:08
basically no dose modulation. So this leads to very high radiation to patients.
24:13
Now, this one is a prospectively triggered but high-pitch helical.
24:19
This is
24:22
a unique scanner specific thing or vendor specific thing in which they have
24:28
dual-source scanners, and this mode is also frequently referred to as
24:34
flash mode. Basically, in this one, what happens is the scanner records the heart
24:39
rate, and in a very fast single sweep, it captures just this portion of the
24:45
cardiac cycle. The problem with this is that if
24:51
you record something about it and the heart rate changes, the scanner cannot
24:55
change, and you just get one shot at it.
24:59
So basically, we use it if the scanner is available.
25:02
This is primarily used in an ideal patient with low regular heart rate.
25:06
If there is any variability, this mode tends
25:11
not to work. Now,
25:14
we talked a little bit about single-source scanner, dual-source scanner.
25:18
I used these terms a few times. The reason this is important is basically
25:24
to get a cardiac image or to see one image on a cardiac scan or to
25:30
acquire that volume of the scan,
25:34
you need to have approximately half rotation around the patient.
25:39
So let's say this is the patient or the heart. You need to have half a rotation.
25:44
Now, in a single-source scanner, which is the most common scanners that we have,
25:49
you need to go through the entire half rotation.
25:52
So for example, if this gantry is rotating at
25:57
240 millisecond or 0.24 seconds, your temporal
26:02
resolution is around 120 because 120 is half of 240.
26:08
Now, if you have two sources or dual-source scanner, your each source can move just
26:14
quarter around or 90 degrees around, and that would each both 90 plus
26:20
90 would make it 180. So the same 250 millisecond rotation gives you
26:26
a temporal resolution of 60 milliseconds.
26:28
So you can actually resolve at much faster heart rates.
26:33
So the way you should think is, depending on what scanner is available in your
26:38
practice, you can meet different heart rate goals.
26:41
So if I have a single-source scanner, like I mentioned, we have a single-source
26:46
scanner. My heart rate goals are to make sure patient has a heart rate preferably
26:51
less than 60, definitely less than 75.
26:54
Now, on a dual-source scanner, are very forgiving.
26:57
You can go at much higher heart rate, and even up to 9,200 I've
27:03
seen reasonably diagnostic scans coming out of dual-source scanners.
27:07
So that's the advantage. Now, what phase of the cardiac cycle you trigger for?
27:13
At higher heart rates, systole is very, very consistent.
27:18
Basically, at slower heart rates end diastole or mid diastole is what we aim for
27:24
because it's the longest quiet phase of the cardiac cycle.
27:30
Now, as the heart rate increases, diastole is the phase that decreases, but
27:36
the systole remains consistent. So that's why at higher heart rates end systolic
27:41
acquisitions are more predictable, so that's what we aim for.
27:47
Same is true for arrhythmia patients, that when you have arrhythmia patients,
27:52
systole is what is more consistent.
27:55
So for example, if you have a patient whose heart rate, who's in AFib, and the
27:59
heart rate moves on the scanner at from 40 to 60 to 80 to 120, the
28:04
most consistent part in those is going to be systole.
28:09
Now, a little bit on why this information is important, because
28:15
you get
28:17
radiation dose based on what you choose. Okay? Radiation dose is an entire topic.
28:23
Just to introduce it here, basically, you get what was the
28:29
tube output, that is CTDI volume. This is the tube output.
28:34
This is dose length product. That means it accounts for the tube output multiplied
28:39
by the length you scan, the z-axis you scan.
28:42
So then if you take this dose length product and multiply it with the chest k
28:47
factor, then that's what you get effective dose.
28:51
So ACR says that the relative radiation level for
28:57
coronary CT is between one to 10 millisievert.
29:00
And for most part, if you see, almost all of the cardiac
29:06
CTs fall under that. The basic modes that we do is
29:11
basically prospective, and that is two to four millisievert.
29:15
If you see this, it says prospective with wide padding.
29:19
Padding just means that instead of acquiring an absolute 70% or
29:25
an absolute 45%, I have a slight buffer.
29:28
I just go from 25 to 45% or 60 to 80%. Padding is that buffer.
29:34
So it adds this, but it basically gives you some more leverage.
29:40
This is the least commonly performed scan, retrospective helical, basically the
29:45
entire cycle without any tube modulation.
29:49
So this is where you should be just a little bit careful when you do that.
29:55
Then you've acquired the scan. How do we visualize it?
29:59
Obviously, these images, more than anything, I just want you to remember the names.
30:04
This is how axial acquisition looks like.
30:07
Basically, this is virtual cross-section imaging through the
30:12
patient's chest. This is how axial images look like.
30:15
We can reconstruct these images in different planes.
30:18
So for example, this is the C view of the RCA.
30:21
These are called as curved planar reformats.
30:25
This basic example is a multiplanar reformat.
30:29
You take axial and you reconstruct an MPR. This is basically how we do it.
30:36
Then you can also see these images as these virtual
30:42
models, and this is called as a volume rendered image.
30:45
You basically have given volume to each structure, and these curvilinear images
30:50
that whenever you see, these are called as curved planar reformat.
30:54
So this is curved planar reformat.
30:57
You can reformat every vessel,
31:00
as a curved planar, it helps you to just visualize the entire vessel in one image.
31:05
But these are the names you should be aware of.
31:10
So one thing to remember that cardiac motion artifacts are usually phase specific.
31:15
That means that, for example, if you are reviewing the coronary CT and you saw an
31:21
image like this, and this is a 75% phase, so this is where at the origin of the LV
31:27
you see pretty tight stenosis.
31:31
Always make sure that if you have available, always look at different phases to see
31:36
if this is just motion. Because in the same patient, if you go and look at it in
31:39
systole, sorry, you go from just 75 to 80% phase, that
31:45
area just completely clears itself. So this is very important.
31:50
So cardiac motion is phase specific, so always
31:54
look for these. Irregular rhythms or high heart rate.
31:59
Obviously, it's implied lower heart rate is better.
32:02
If you have very high heart rate, it's one justification to do retrospective
32:07
gating.
32:08
But one important thing I want you to take home is faster heart rates you want to
32:14
acquire in systole. Obviously, go for the highest temporal resolution.
32:19
Whenever heart rate is high, avoid any other artifacts.
32:23
Make sure your contrast delivery is perfect, your signal-to-noise ratio is perfect,
32:27
patient is not bleeding. And
32:31
the likelihood of success based on all factors.
32:33
If there's a low pretest probability, low calcium score, and for broader graft
32:39
potency, usually it's okay, but patient who has very high heart rate, high BMI, and
32:44
you're looking at multiple small scans, it's going to be challenging to give out
32:49
their answer. So basically, understanding nuances
32:55
on how to do this is very, very important.
32:58
And also understanding any kind of arrhythmia management
33:04
that you can do by ECG editing is very important.
33:07
So for example, this is one case in which, if you see, this was a
33:12
patient, came in with high heart rate.
33:15
In diastole, this is how the images look like.
33:18
Really cannot do or diagnose and rather than report this as non-diagnostic,
33:24
just look in systole.
33:27
You can always, not always, you can most times find a phase in which you can
33:32
diagnose things. Same thing. This was a patient, we did it on one of-- so basically
33:38
it's an example from one of the older generation scanners.
33:42
And if you see, there is slab artifacts, and you can see how much of these slab
33:47
artifacts happen when patients have slightly irregular or any heart rate,
33:55
like varying RR interval. So this was a patient with arrhythmia, and this scan
33:59
practically looked non-diagnostic.
34:02
And if you look at the patient's
34:05
ECG trace in the scanner, and you can see there's a lot of, basically
34:10
the
34:14
patient recorded a lot of PVCs as
34:19
QRS complex. So all you have to do is delete some of these
34:25
abnormal heartbeats and just label the correct QRS and label the PVCs.
34:31
And as soon as you do some ECG editing, the same case went from being completely
34:37
non-diagnostic to having a fully diagnostic scan.
34:41
So, this is where I think, and this is called as ECG editing, and some of the
34:47
advanced applications function, obviously, for ECG editing is typically possible
34:53
in retrospectively gated scans only.
34:56
For prospective scans, one thing that we really focus on is arrhythmia
35:02
rejection. And in these patients, what happens is,
35:07
the scanner records the expected heart rate.
35:10
Now, if the expected heart rate is different by a certain
35:14
percentage, so for example, let's say 1,000 milliseconds is the expected
35:21
RR interval, or basically it's
35:24
60 beats per minute. Now you suddenly start acquiring and patient
35:29
had a PAC, and then the RR interval went to 700, it would
35:35
stop. It would not scan that heart rate, and it would move to the next heart rate.
35:39
And as soon as the heart rate is within 10% of what was the expected heart rate,
35:44
that means it went from, let's say, 950 to 1,050, it would acquire
35:50
the scan. So these are called as arrhythmia rejection algorithms.
35:54
These are very useful and should be turned on.
35:57
Another thing is atrial fibrillation, one of the most common scenario that you will
36:03
see. For these, we do prospectively gated scans with end
36:09
systolic reconstruction. So one very important tip for AFib patients is
36:15
that do not do percentage RR based reconstruction.
36:19
Just do a millisecond after R wave.
36:22
The reason for this is that, for example, if you were doing percentage-based
36:26
reconstruction, and this is the patient's-
36:30
Our basically ECG rhythm. 50% here is going to be, let's say all of
36:36
these initial ones are 60 beats per minute or 1,000 milliseconds.
36:40
So 50% is 500. This is approximately 500.
36:44
Now, this patient's heart rate varied, and now 50% is 800 because it became
36:50
1,600 milliseconds. Then it came back, then it became slower, and now the 50% is
36:56
300 milliseconds. That's the problem with percentage-based reconstruction.
37:00
Whereas if you use millisecond-based reconstruction, it basically moves.
37:05
It's 500 milliseconds after the previous R wave.
37:09
So every time it's going to be consistent, even if the RR interval keeps varying.
37:15
So this is one very important point.
37:18
You can really save a scan in a patient with AFib by using these millisecond-based
37:23
reconstruction. Same thing, moderately obese patients.
37:27
The problem with them is how much radiation that can go in.
37:31
What you can do is, obviously, you want the maximum photon
37:36
penetration, so at least use 120 kV.
37:42
And with that kV, maximize your tube mA.
37:46
These are the two things that you can do.
37:51
Also, whenever there is an obese patient, I always focus on how important it is
37:56
heart rate control. When you have an expectation of a slightly lower quality, you
38:02
cannot compromise on other things. Okay?
38:05
So other things is you can use a soft filter or a soft kernel.
38:10
Do not use, let's say, a lung kernel.
38:12
Otherwise, obviously, unless they do the stent, then you have to.
38:16
You can use higher degree of iterative reconstruction, like deep learning
38:20
reconstructions. So with this, in our experience, this is a, for example, a
38:25
patient whose BMI was close to 51, and we can get very
38:31
high-quality imaging, even with very high BMI.
38:37
You can also use, as I said, a higher degree of iterative
38:42
reconstruction. Not every iterative reconstruction I would suggest, but certain are
38:48
very good. This is one of the exercises we did last year, and we moved to higher
38:54
degree, higher strength of deep learning iterative reconstruction at our scanner.
39:00
And it's something that we published on is how it really does
39:06
not affect image texture or our ability to characterize stenosis, but just
39:12
decrease the noise. And noise is this graininess that you see on image, and see how
39:16
it is decreasing without changing our confidence in diagnosing this
39:22
area of tobacco stenosis. So in stents or high calcium, the
39:28
thing that we do is obviously you want to have good amount of radiation going in.
39:33
So, 100 to 120 kVp have good amount of mA going in,
39:39
and beyond that, you need slightly sharper kernels for reconstruction.
39:45
So these are the things
39:48
we focus on. And for stents,
39:51
our technical learning is that more than three-millimeter stents are okay.
39:58
Obviously, with photon counting, I think that's going to be redefined.
40:02
I have a feeling that we are able to see lumen of even smaller stents.
40:08
Okay. One very big tip is while interpreting these exams,
40:14
always widen the window and level, because that would decrease your blooming
40:20
through this. So that's very important when you are reading these scans.
40:24
This is just an example of our experience.
40:27
These were our conventional energy integrating detector scanner, and we basically,
40:33
the same patient got a scan on a photon counting scanner, and you can see how well
40:38
you can see the lumen. And you tend to underestimate calcium.
40:43
You basically characterize the stenosis much better on these photon
40:49
counting scans. Same has been our experience with smaller stents.
40:53
Like I said, this was a 2.5-millimeter stent in the left circumflex coronary
40:58
artery, and very easily we can diagnose things with this.
41:02
So,
41:04
I'll just slightly quickly go through a few things like this on applications of
41:10
how we are using cardiac CTs, not just beyond evaluating
41:16
stenosis and atherosclerosis. This was a 25-year-old patient who was having chest
41:20
pain evaluation, and these are the things in which cardiac CT is very good and very
41:26
important. This patient, if you see, has a coronary anomaly in which you see
41:32
a proximal part of the coronary artery.
41:35
The right coronary artery is coming from the left, and then it has a narrow
41:40
slit-like, or it's really seen very well in these volume rendered images.
41:45
So this is the basically left sinus of Valsalva, which gives out the left
41:51
coronary artery, and from the same is coming the right coronary artery with acute
41:55
angle as well as slit-like origin.
41:57
So basically what we have is an anomalous right coronary artery from the left cusp
42:03
with an inter-arterial course. Now, a few important things to highlight in these
42:08
patients is obviously the intramural segment of these.
42:12
That means if there is a segment that is inside the wall of the aorta, that's very
42:18
critical. And for that, we look at this slit-like
42:22
origin as well as lack of fat around it.
42:26
And not only that, coronary imaging also allows you to do endoluminal visualization
42:31
that sometimes can help and look at the lumen originating.
42:36
So this is in the same patient. This is the left coronary artery with very
42:40
round-looking lumen. There is the right coronary artery, has a slit-like lumen.
42:44
So you can give more perspective to the surgeon as well.
42:49
This is one excellent paper I always recommend people to read for anomalous
42:54
coronary arteries types and how we differentiate different types of coronary artery
43:00
origins. And this is one of my favorite figure out of this, because
43:05
inter-arterial course and subpulmonic or transseptal course gets confused very
43:10
frequently. And this differentiation is if you have anomalous coronary artery at
43:15
the level of the pulmonary valve, it is inter-arterial.
43:20
And if the same very similar course, but if it is at below the level of the
43:25
pulmonic valve and just going through the septum, then that's subpulmonic or
43:30
transseptal course.
43:33
This is just another example of how coronary imaging is useful and plays a
43:38
complementary role. This is a patient with chest pain, shortness of breath, and had
43:43
this echo. And on this echo, what was read is that this patient has very dilated
43:48
left coronary artery. And we suspected based on this echo and Doppler that there's
43:53
probably a fistula of why this is dilated.
43:57
But these are normal origin dilated coronaries, suspicion for fistula.
44:03
And we got this coronary exam thinking that we are going to find fistula.
44:07
And,
44:09
I'm sorry this is slightly fast. But focus on the origin of these coronaries.
44:14
And if you see this image, what you will see is that the
44:20
right coronary artery is originating from the aorta, but the left is
44:26
originating from the pulmonary artery.
44:29
So this is basically-- And you can understand why on echo it looked like it is
44:34
coming from the correct cusp because on this axial image at the level of the sinus
44:39
of Valsalva, that's how it looked. So what we were seeing was this image on echo.
44:45
What we did not see is this origin coming from here.
44:49
So that's why coronary CT is just so useful.
44:53
And not only that, this is another example in which this patient did not have
44:57
access to care, and the patient came in with a normal chest X-ray.
45:01
And this patient had
45:05
basically an infarct in the left circumflex distribution with a large
45:11
aneurysm, as well as presence of thrombus in that large aneurysm.
45:16
And just with cardiac CT, you can diagnose a lot of things.
45:19
And incidentally, on this patient, this patient also had an imperial
45:25
sinus venosus ASD. That means the IVC was not just going in the right atrium,
45:32
it was also connecting with the left atrium. So it was straddling together.
45:37
So Everything structural you can get from this test, and that's
45:43
why we see so much use. Now, very important area is also chest pain.
45:48
Like chest pain,
45:50
it's like preaching to the crowd. If you are here, I know you are interested in
45:54
this topic. So it's basically the second most common cause of ER visits, with
45:58
nearly eight million visits a year.
46:01
And what we have learned about data is that CCTA is,
46:07
if you define significant stenosis as more than 50%, CCTA is very
46:13
sensitive and the only thing where if you look at the data was
46:19
the specificity was low, especially when you look at the functionally significant
46:23
disease as compared to invasive FFR.
46:26
So this was the initial data which said, okay, CCTA has good sensitivity.
46:31
It has excellent negative predictive value.
46:33
That means when it is normal, the chances that this patient has anything wrong in
46:39
the coronary arteries is very, very low.
46:42
What it did not have was good specificity versus for functionally
46:47
significant or real IFR-positive disease.
46:53
But as we know, our
46:57
technology is improving. As we are going from a 16-slice scanner to now more than
47:02
128-slice scanners, our temporal resolution is reaching
47:08
around 60 millisecond now, and we are
47:13
requiring less contrast, we are giving less radiation, and we are having less
47:16
artifact. I already went through this.
47:19
Basically, this is showing how dual source scanners really are pushing the
47:23
boundaries of how fast you can image these patients.
47:27
And what we are learning is that more we see, so this is the data from CONFIRM
47:34
1 registry. Now, right now, we have CONFIRM 2 data that is coming out.
47:38
And what we are seeing is more disease versus outcome.
47:44
So survival probability when the CT is normal versus whenever there is a
47:48
three-vessel disease or left main disease is really, really bad.
47:51
But as you see, more disease, bad outcomes.
47:55
And not only that, this was shown in multiple pathway, in
48:01
multiple studies that even patients who have non-obstructive disease versus no
48:07
disease, the survival probability just with presence of any disease, even if it is
48:13
non-obstructive, is significantly decreased.
48:15
And we can see that one vessel versus two vessel versus three vessel disease, how
48:21
hazard ratio increases in these patients.
48:27
We know that, as we said, CCTA has very good negative predictive value.
48:32
So as we go and we have a negative scan, this really improves.
48:38
Basically, if you look at the ED studies, that if you have
48:44
a negative or a normal CCTA, the annual event rate with CT
48:50
is extremely, extremely low. Basically, it's 0.3%.
48:54
Versus if you compare it with SPECT, another test that is used for ischemia
48:59
evaluation, the event rate is higher.
49:02
And what we are learning with the SCOT-HEART registry, which I think really gives
49:08
insight into a lot of things because this was the initial SCOT-HEART paper that
49:13
talked about the five-year data that people who came in with chest pain and got CT,
49:20
just getting CT versus the standard of care, really the incident rate was
49:25
less. So same thing in PROMISE, we saw that patients
49:31
who got CT versus they were triaged with other modalities, the event
49:37
rate was much lower in the CT arm. So the question is: What is leading
49:42
to lower event rate and lower
49:47
rate of MACE in these patients, both for coronary death as well as
49:53
non-fatal myocardial infarction is basically what was shown was patients who got
49:59
CT, we were able to see even their non-obstructive disease as well as entire extent
50:04
of disease. So these patients were more on preventive therapies or when needed or
50:10
when indicated on aspirin, and that drives a lot of positive
50:15
outcomes in long-term. We obviously, when we report a CT, we follow
50:21
what is called as SCIRADS 2, in which we give a certain
50:27
basically grading of stenosis, which is basically no, minimal, mild, moderate,
50:31
severe. And
50:34
the advantage of using these standard techniques is that
50:41
this is where if you see, let's say you have a 50-year-old female that came
50:47
with atypical chest pain. Now, the question is, your CT here
50:52
is a normal coronary CTA. This patient has some non-obstructive disease
50:59
in three vessels. So all three vessels have disease, and this patient has
51:03
non-obstructive disease, but this plaque has highest features with positive
51:08
remodeling and low attenuation plaque.
51:10
But if you see, the nuclear scan or the SPECT exam in all
51:16
three scenarios would have been normal.
51:19
But obviously, the care of these patients does not have to be same.
51:24
This has to be different. And this is where CT really adds to it, and it has been
51:30
shown to be safe. It has been shown to be very efficient. It is fast.
51:36
You see ED turnaround times are much faster with use of CT.
51:42
And you can not only give stenosis, you can even give native plaque features out of
51:47
which we mainly talk of positive remodeling, low attenuation plaque,
51:53
neck-in-ring appearance, and spotty calcification.
51:56
These are the four highest features we frequently talk about in these, and these
52:01
are associated with basically also poor outcomes.
52:07
So for example, even if one feature is positive versus two feature positive, more
52:13
the bad features, the worse the outcome.
52:15
And there is so many studies that are coming out.
52:18
This is basically from SCOT-HEART trial that just total presence of low attenuation
52:23
plaque, which is a slightly counterintuitive in my head.
52:26
I always think of low attenuation plaque in a certain plaque as a risk of
52:32
rupture. But they just took total coronary volume, low
52:38
attenuation plaque, and if it was more than 4%, just look at how high the event
52:43
rate was versus people who had less than 4%.
52:46
So very useful data, and this is how plaque kind of
52:52
comes in the reports that you can quantify absolute amount of total plaque,
52:58
non-calcified plaque, calcified plaque.
53:01
And there is so much data coming out on this and how total plaque burden
53:07
and not just stenosis is the main predictor and risk of cardiovascular events and
53:12
death. Obviously, stenosis is a driver for patient symptoms, but plaque
53:17
burden is a very big driver of cardiovascular events as well as
53:23
death. So,
53:28
I would just briefly add to this that we also have CT is just seen as an anatomical
53:34
test, but we have an ability to add a functional test to this, and that
53:39
functional testing is what is called as a CTFFR.
53:43
And it's a generic term, I would say.
53:46
The reason I say CTFFR is a generic term is because it can be done with
53:52
computational flow dynamics, and there are multiple companies that have used
53:56
different approaches. One of the common one approach used is computational flow
54:01
dynamics, and this is basically an image based on this.
54:06
So what this image is showing is you are seeing stenosis.
54:10
Now, can you predict the functional significance of the stenosis?
54:15
And this is prediction.
54:18
Basically not absolute, so please do not take it as, oh, this shows less than
54:24
0.8,
54:25
and this has to be significant. So basically, CTFFR value is more than
54:31
or equal to 0.8 is normal. Less than 0.75 are abnormal, and
54:36
0.75 to 0.79 are kind of intermediate range.
54:41
But these are The CT-derived FFR. Now they are more than
54:48
computational flow dynamics models.
54:50
So please, there are AI-based models these days.
54:53
There are so many models that are now FDA approved, and that is not just the only
54:58
way to do computational flow dynamics.
55:01
But I just want to highlight how great that data is.
55:06
So basically, this is from one of the study that was done.
55:11
The end was not very high, but for stable chest pain, what they did was they used
55:17
how different modalities compare.
55:20
CTFFR with CCTA versus CCTA alone with no CTFFR
55:26
versus PET, versus SPECT. And the comparison is gold standard, that is invasive
55:31
FFR. And if you see these colors, what you will see is that fractional flow reserve
55:36
with CTFFR has had the best area under curve of 0.94,
55:42
followed by PET, followed by coronary CT alone, and then SPECT.
55:48
So this is basically how these things are being reflected.
55:52
And I want to add that this is data that recently came out
55:58
from 10 years from SCOT-HEART. I briefly showed you five-year SCOT-HEART.
56:03
And in long term, this data kind of continues to show the same.
56:09
If you go from five to 10, the benefit of triaging our patients through
56:15
CCTA continue to hold over time. And not only that,
56:21
you can see that these are non-fatal MI
56:26
in this group is basically lower, although the coronary revascularization
56:32
is similar. So it did not lead to multiple more coronary vascularization and yet
56:38
lead to less things. And the main thing it did it is that by having more
56:44
proportion of patients on preventive medication.
56:48
And if you see in all the years, this benefit kind of continues to go.
56:53
It was there at year five. It continues in year 10.
56:56
That's why this has become part of our guidelines as well as it has become part of
57:02
all of our
57:04
current chest pain guidelines and preventive guidelines.
57:07
With that, I would like to stop and answer these questions because I can go on for
57:13
another many, many minutes. So let me.
57:17
Thank you so much for this lecture. Yeah, there's a bunch of questions in the Q&A.
57:22
I say pick a couple since we're at time and yeah.
57:25
Okay.
57:27
What software do you use for plaque analysis?
57:32
So we have currently tried HeartFlow, and we tried it
57:38
as a part of our partnership at the SITE registry.
57:42
So that's the one that I have experience with.
57:45
But I've heard very good things about other softwares as well.
57:53
Have you begun looking at coronary artery periarterial fat attenuation
57:58
changes? I understand this may be a useful indicator of patients at risk for future
58:02
vascular events.
58:04
Absolutely. The data on PCAT, pericoronary adipose tissue or
58:12
FAI is increasing, and there's a lot of data that is suggesting that you can use it
58:18
as a marker of future events. I do have some concerns
58:24
about these techniques, is that, one, the study that showed that minus
58:30
70 is the magic number was only used in patients with stable chest pain.
58:36
And that, too, they characterize it along the right coronary artery.
58:40
Not sure if we can translate the same data to every coronary artery or every
58:44
patient scenario. I also think that the attenuation of these
58:49
fat is very dependent on protocol.
58:52
Like higher kV would change how the attenuation changes.
58:57
So I think there's a slight need to homogenize our protocols to make sure that
59:02
these are interoperable and consistent, so that the same patient goes to two
59:08
different sites, gets the same quantitative number.
59:11
But I do think as a
59:16
basic technique, I think it has a lot of value.
59:22
I was able to find my QA, so I can go from the top.
59:27
It says, right heart, we trigger off right pulmonary artery.
59:32
What is the attenuation that needs to be achieved?
59:34
So again, depends on what you are trying to achieve.
59:37
For our tricuspid imaging,
59:40
when we do right heart triggers, we usually rely on 150 Hounsfield units, and we
59:46
do give slightly diluted contrast, and that usually gives good imaging.
59:51
So I would say 150 is a good trigger to have for right heart evaluation.
59:58
Then it says, can you please re-explain how you scan protocol when you want
60:04
to concentrate contrast in the right ventricle, when you want to concentrate in
60:08
both ventricles? Okay. So whenever we are trying to do a right ventricle only
60:13
study, usually what I do is that I
60:17
use a biphasic kind of a protocol in which we take a prediluted contrast.
60:23
It's not a full
60:25
contrast, but it's a prediluted contrast.
60:28
We trigger off the main pulmonary artery at 150 Hounsfield unit, and there is a
60:33
saline chase that keeps coming after this. So that's how we do it.
60:37
Now for both ventricles, we don't use a prediluted contrast. We use that saline.
60:44
Sorry. We give concentrated contrast, then contrast saline mix, then saline.
60:50
And in that, we trigger off the aorta, descending thoracic aorta, and that is how
60:54
we achieve basic
60:57
enhancement. So the goal is the first bolus of the contrast that is coming in is
61:02
your bolus for the left heart. The contrast mix that is coming in, contrast saline
61:07
mix, that's your contrast for the right heart.
61:10
That's the goal of doing that kind of a study.
61:15
What are our options for dealing with motion after the scan has been performed?
61:19
That's a very good question. I think that I approach this two ways.
61:24
One, looking at different phases.
61:27
So obviously you want to be looking at different phases.
61:30
Now, also, a lot of scanners have motion correction algorithms.
61:37
I can tell you for single-source scanners, most because dual source do not have
61:42
these issues, so they have not really focused on this.
61:45
But at our site, I have GE scanners as well as Canon scanners, and both
61:51
have AI-based motion correction.
61:54
They really help in motion correction even at higher heart rates.
61:57
So I would suggest if you have a scanner, always turning on your motion correction.
62:03
In GE, it is called Snapshot Freeze.
62:07
I'm blanking on the name on Canon, but I think it is basically it's called some
62:13
form of a freeze or something, motion correction.
62:16
So just look at these options for your scanner.
62:23
Another one says, how do you change your protocol for grafts and stents?
62:27
Contrast volume sequential versus high-pitch helical reconstruction.
62:32
Great question. Grafts, you want to go higher, make sure that you always
62:37
include your Lima origin. Obviously, most commonly Lima is used.
62:43
So first is acquisition. Second is contrast volume.
62:47
We usually just increase it by 25 ML.
62:50
We do not change a lot in terms of contrast volume.
62:54
We still increase a little bit more because you are covering more territory.
62:58
You are covering more heart rates.
63:01
We
63:02
don't do high-pitch helical. High-pitch helical is only done for very young, ideal
63:08
heart rate, ideal patient. We rely very heavily in our
63:14
institution with sequential scanning.
63:17
I would say in bypass evaluation, either sequential or
63:22
retrospective helical with dose modulation are usually the right answers.
63:28
In terms of reconstructions, only thing I would change with only with stents, I
63:33
would not do anything different with grafts.
63:35
But with stent, you want basically a sharper kernel, which is either a lung or
63:41
a bone kernel. So just reconstruct one sharper kernel, which is bone or lung, and
63:48
with that reconstruction, you should be reading the exam.
63:51
So that's the one difference we do in stent patients.
63:56
How do you manage a patient with no
63:59
obstructive blood but high calcific blood burden?
64:04
I think this is a little bit of a question that
64:10
may not pertain much to cardiac CT, but managing the patient.
64:15
And I don't want to pretend to be the right person to answer this the
64:21
best way. But I do think whether if there is high calcific burden,
64:27
and there's no obstructive blood, the only binary difference I think is that, okay,
64:32
this patient does not go for CAT. But this patient goes through all the medical
64:37
management that this patient needs based on the risk factors as well as presence of
64:42
plaque burden. Now, obviously, there are guidelines based on, in the preventive
64:47
guidelines, they are based on what percentile the patient is on and
64:53
what is the absolute calcium score.
64:56
And based on that, it tells you how aggressive you need to be on
65:01
preventive therapy.
65:04
The guidelines on plaque quantification, absolute plaque quantification, not
65:08
calcium score quantification, but absolute plaque quantification are still a little
65:13
early. Although many registries and many clinical trials have tried to suggest,
65:19
oh, beyond this certain, you should add a PCSK9 inhibitor.
65:24
Beyond this, you add cholestyramine.
65:26
I don't know if they have penetrated
65:30
the clinical practice. But at the same time, as a radiologist, I really am not the
65:36
right person to fully answer this question.
65:41
Well, thank you so much for getting through all those questions and for your
65:44
lecture. Appreciate that.
65:46
Thank you. Thank you, Ashley. Thank you everyone for attending.
65:49
Yes. Thank you for everyone for participating and asking such great questions.
65:54
You can access the recording of today's noon conference and all our previous ones
65:58
by creating a free account. We'll also email out a link to the replay later today.
66:02
Be sure to join us next week on Tuesday, September 15th at 12:00 PM Eastern, where
66:06
Dr. Dorsaman will deliver a lecture entitled Introduction to Brain MRS.
66:11
You can register for that at medallia.com.
66:13
Follow us on social media for updates on future noon conferences.
66:16
Thanks again, and have a great day.